基于小波变换和ARMA-LSSVM的忙时话务量预测

Forecasting of busy telephone traffic based on wavelet transform and ARMA-LSSVM

为提高受多种因素影响的话务量数据的预测精度和稳定性,提出一种考虑多因素影响的基于小波变换和自回归滑动平均(ARMA)-最小二乘支持向量机(LSSVM)的话务量组合预测模型。对忙时话务量数据进行相关性分析,得出影响话务量的重要因子;利用小波变换对数据进行分解和重构,得到低频分量和高频分量;将低频分量输入ARMA模型进行预测,将高频分量和话务量重要影响因子输入粒子群算法优化的LSSVM模型进行预测,将两组预测结果合成。实验结果表明,该模型进一步提高了预测精度和稳定性。

To improve the prediction accuracy and stability of telephone traffic which are influenced by multiple factors,a combined forecasting model was proposed which took the influence of multiple factors into consideration and combined wavelet transform,auto regressive and moving average（ARMA）model and least squares support vector machines（LSSVM）model.The correlation analysis was firstly applied to the busy telephone traffic data to obtain the key factors which influenced the busy telephone traffic.Then the wavelet transform was used to decompose and reconstruct the telephone traffic data to get low-frequency and high-frequency components.The low-frequency component was loaded into ARMA model to predict,while the high-frequency component and the obtained key factors were loaded into LSSVM model that was optimized by the particle swarm optimization（PSO）to predict.Finally the forecasting result was achieved by the superposition of predictive values.The simulation results show that the proposed model improves the prediction accuracy and stability.